J Clin Pathol 1995;48:535-538

Desmin expression in rhabdomyosarcoma:influence of the desmin clone andimmunohistochemical method

L Pollock, D Rampling, S E Greenwald, M Malone

AbstractAim-To determine which, if any, of fivecommercially available desmin clones ismost reliable at labelling desmin filamentsand whether the enhanced polymer onestep (EPOS) method of labelling is ofany advantage in the routine diagnosticlaboratory.Methods-Thirty four rhabdomyosar-comas from the files at The Hospital forSick Children, Great Ormond Street, Lon-don, were studied. Four different desminclones, DE-R-11, D33, DE-U-10, andPDE, were applied to each using the con-ventional extravidin biotin peroxidasemethod. The D33 clone was also appliedusing the EPOS method.Results-The EPOS method incor-porating D33 persistently scored morecells as desmin positive and was positive infour cases which were negative on stainingwith the other clones.Conclusions-The D33 desmin clone usedwith the EPOS method is more reliable foridentifying desmin filaments in tumoursthan other desmin antibodies tested.Different desmin clones using a routinetechnique label different rhabdomyo-sarcoma cells and therefore it is justifiableto use more than one clone.(7 Clin Pathol 1995;48:535-538)

Keywords: Desmin, rhabdomyosarcoma, immuno-histochemistry.

Department ofHistopathology,Great Ormond StreetHospital for SickChildren NHS Trust,LondonL PollockD RamplingM Malone

Department ofHistopathology,The Royal LondonHospital MedicalCollege, LondonS E Greenwald

Correspondence to:Dr L Pollock,Department ofHistopathology, King'sCollege Hospital,London SE5 9RS.Accepted for publication13 October 1994

Tumours of skeletal and smooth muscle maypose difficulties in diagnosis. Desmin is a pro-tein found in cells of muscle origin' and desminantibodies are now well established as the bestroutine marker for identifying rhabdomyo-sarcomas.23 In evaluating the effectivenessof the new enhanced polymer one step (EPOS)immunohistochemical method in a rhab-domyosarcoma case, we noted that the numberof positively labelled cells and the intensity ofstaining varied enormously between the EPOSmethod and the usual extravidin biotin per-

oxidase technique. We decided to investigatethis phenomenon further by comparing theeffectiveness of different desmin clones and thetwo different methods of applying the antibody.Desmin is one of the family of intermediate

filaments composed of a non-helical headpiece,an oa helical middle domain and a C terminaltailpiece.' Antibodies have been raised againstdifferent portions of the protein, each being

identified by its clone code-for example, DER11 and D33.5 An ever increasing array ofdifferent antibody clones is available com-mercially for routine immunohistochemicaluse.6The EPOS technique is a new method of

immunohistochemical labelling. It is a directstaining technique made possible by the linkingof the primary antibody to horseradish per-oxidase (HRP) by a polymer which then com-plexes with a greater number of enzymemolecules than the antibody alone, thus affect-ing amplification without the need for multi-layering. The striking advantage of the EPOSsystem is in time saved and the fewer stepsrequired. There is no need for titration, re-ducing the risk of errors, and a result can beobtained within a few hours.To evaluate which desmin clone and which

labelling method would be most useful in ourlaboratory, we compared four different desminclones and with one clone we used the EPOSmethod as well as the routine extravidin biotintechnique. As parts of the desmin moleculeshare extensive homology with other in-termediate filaments, we also labelled 50 casesof non-rhabdomyosarcoma childhood tumourswith the same desmin clones to assess thepossibility of cross-reactivity.

MethodsThirty four cases of rhabdomyosarcoma (26embryonal and seven alveolar), confirmed byelectron microscopy, and 10 cases each ofneuroblastoma, lymphoma, Wilm's tumour,primitive neuroectodermal tumour (PNET),and retinoblastoma were retrieved from thefiles of Great Ormond Street Hospital for SickChildren, London. All of the tumours had beenfixed in formalin and embedded in paraffinwax. Sections (3 gm) were cut, dewaxed andrehydrated. The slides were labelled with fourdifferent desmin clones using a standard ex-travidin biotin peroxidase technique. En-dogenous peroxidase was blocked in all slidesby incubating with 10% hydrogen peroxide for20 minutes. Details of digestion, incubationtimes and antibody source are given in table 1.Slides were then incubated with the appropriatesecond antibody. For the mouse monoclonal,biotinylated goat antimouse at a 1 in 50 dilutionfor one hour and for the rabbit polyclonal,biotinylated goat antirabbit at a 1 in 200 di-lution for 30 minutes.The slides were washed in phosphate

buffered saline (PBS) and incubated with either

535

on Novem

ber 28, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.48.6.535 on 1 June 1995. D

ownloaded from

Pollock, Rampling, Greenwald, Malone

U _- / t , = .r

7. S __ _ _z | | S | 1 i. 4- I w _"

:, -,. Iv\.-

.

\ N.K.K.:\ N'.

'IX....Figure 1 Number of cases which fell into each of the five different percentage labellinggroups for each of the 66 observations per antibody: 0= 0%; 1 = 1-5%; 2= 6-25%; 3=26-45%; 4=46-65%; 5= >65%. .,.f 0

if.'ir

A

,XIdf ,r i tX. %* -i

#N4

Table 1 Digestion, dilution and source of the fourdifferent antibodies

Antibody Type Protease Dilution Source

DERlI mm 5mins 1 in 75 DakoD33 mm none 1 in 50 DakoDE-U-10 mm 5mins 1 in 150 SigmaPDE rp none 1 in 50 Europath

Dako, Glostrup, Denmark; Sigma, Poole, Dorset, UK;Europath, Cornwall, UK.mm =mouse monoclonal; rp =rabbit polyclonal.

mouse monoclonal antiserum (diluted 1 in 50)for one hour or for the PDE clone, rabbitpolyclonal antiserum (diluted 1 in 200) for30 minutes. The peroxidase activity was de-veloped with 3'3-diaminobenzidine (DAB) for10 minutes. All incubations were carried outat room temperature. The slides were thencounterstained with Mayer's haematoxylin.Negative controls omitted the primary anti-body; appropriate positive controls were run inparallel.For the D33 EPOS method, the slides were

dewaxed and rehydrated and then incubatedfor 20 minutes with 100% hydrogen peroxidein distilled water, rinsed and placed in PBS forfive minutes. The slides were then incubatedat room temperature with EPOS antidesmin/HRP (Dako, Glostrup, Denmark) for 60 min-utes, rinsed in PBS for five minutes and in-cubated with DAB for 10 minutes.

All of the slides were examined by two in-dependent observers and coded such that theobservers did not know which antibody hadbeen applied. The slides were scored accordingto percentage of tumour cells labelled.

*

tI4 4

Figure 2 D33 EPOS labelling of the same field in the same rhabdomin fig 3 to compare the number of tumour cells labelled.

4 ~ ~ 4~

''p

4f.' v ** *'e

t* b *9* ;J*-t r

'. * I A,.|I

*'"'XI*

,4'̂ "S#' ~ '

1. .,.*% +e

* "A..t

C.:.'"" Results'jJ The rhabdomyosarcoma cases were scored by

each independent observer according to thenumber of cells labelled. Intensity of labellingwas not scored separately as there was noquantitative relevance in this study. A cell wasscored if it was clearly positively labelled re-gardless of depth of colour. To quantify the

yosarcoma case as percentage of cells labelled, the slides weregrouped and a score applied to each group asshown in table 2.The scores were then added for each clone

and plotted onto a bar chart (fig 1). Using thex2 test, there was no statistically significantinterobserver error and so the two results couldbe added together. One case had no tumourleft in the block.Two cases which behaved as rhab-

domyosarcoma clinically and which had fea-tures of rhabdomyosarcoma on electronmicroscopy were negative on staining with alldesmin clones by both methods; one was analveolar rhabdomyosarcoma and the other em-

It

.

04

V

:v4*

Figure 3 The samefield as shown infig 2 labelled with desmin clone DE-U-10.

Extravidin biotin peroxidase method.

bryonal.The EPOS D33 label produced a positive

result in four cases when all the other antibodieswere negative. Two antibodies, DE-U-10 andD33 (routine method) were negative whenDERI 1, PDE and D33 were positive on EPOSin four and two cases, respectively.The cumulative results per desmin clone for

the rhabdomyosarcoma cases (33 cases each

536

i I

0

m

on Novem

ber 28, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.48.6.535 on 1 June 1995. D

ownloaded from

Desmin expression in rhabdomyosarcoma

Table 2 Scoing for each percentage group

Per cent labeled Score

0 01-5 16-25 2

26-45 346-65 4>65 5

with a result from two observers making 66observations in total) (y axis) are shown in fig1. D33 EPOS had the lowest score for no

labelling at all in a case-that is, the white area

at the base of the chart. Between the other fourantibodies, although DE-U-1 0 is marginallyhigher for no positive labelling, the differenceis not significant (X2 test). Similarly, D33 EPOShas the highest number of high scoring ob-servations (dark block at the top of the chart)and if the two top blocks are considered-thatis, any slide with more than 46% labelled cells,D33 EPOS picks up twice as many slides as

the next highest scorer which is D33 using theroutine method. Figure 2 shows one rhab-domyosarcoma case labelled by D33 EPOSand fig 3 shows the same field labelled by DE-U-10. The comparison clearly demonstratesthat far more rhabdomyosarcoma cells are

labelled with the D33 EPOS method.None of the 50 cases of childhood non-

rhabdomyosarcoma tumours were positive on

staining with any of the desmin clones. Noneof the 10 Wilm's tumours showed rhab-domyoblastic differentiation.

DiscussionRhabdomyosarcoma is a rare tumour but ac-

counts for 5% of all paediatric malignant tu-mours.' Distinguishing rhabdomyosarcomafrom other childhood small round cell tu-mours-for example, lymphoma, neuro-

blastoma and PNET, may be difficult but isimportant as the treatment protocols and prog-noses vary between the different diseases.8

Age, site, clinical presentation, and imagingcontribute to the diagnosis but histological con-

firmation is necessary. The development ofdesmin antibodies has made a significant con-

tribution to easing this diagnostic process butthey are not foolproof.9"-Desmin is a member of the intermediate

filament family of proteins and was initiallythought to be specific to skeletal, cardiac andsmooth muscle. More recently, this has beenshown not to be absolute as myofibroblastsmay also express desmin.'2 Desmin antibodiesbecame available for use as immuno-histochemical markers in the early 1980s andsince then have become well established as

an aid in the diagnosis of tumours of muscleorigin.'3 Desmin antibodies have proliferatedsuch that there are now a large number ofcommercially available clones. Although thesensitivity and specificity ofdesmin as a markerof rhabdomyosarcoma remains high, other tu-mours can also be labelled with desmin anti-bodies, notably glial tumours'4 and, of more

relevance to paediatric tumours, there is one

report of positive staining with desmin anti-bodies in childhood neuroectodermal tu-mours. 15Most biopsy specimens of small round cell

tumours of childhood are small and thereforethe antibody which reliably labels most cells isthe most valuable for diagnosis. From the res-ults of our study, although there was no sig-nificant difference in the number of positivelylabelled cases with the routine extravidin biotinperoxidase technique between the differentantibodies, there was a difference in the groupsof cells that the different desmin cloneslabelled-that is, two different clones may bothbe positive but due to labelling of differentrhabdomyosarcoma cells. This means that in asmall biopsy specimen there is a greater chanceof identifying desmin positive cells if twodifferent clones are used.The most striking difference in labelling was

that between the routine method and the EPOSmethod. Four of the cases which were negativeon staining with all antibodies by the routinemethod were positive with D33 EPOS, animportant finding as a desmin negative resultmay suggest a diagnosis other than rhab-domyosarcoma.As fig 1 clearly shows, D33 EPOS had the

least number of negatively labelled cases andthe greatest number ofpositively labelled cases.This difference was sustained not only againstthe different desmin clones but also against thesame D33 clone using the routine labellingmethod. Therefore, the clone is not more sens-itive but the technique permits greater antigenretrieval. To test the hypothesis that D33 EPOSis significantly better at labelling desmin thanthe other groups, we applied the y2 test whichgave p<0-001.DE-U-10 and D33 were the only two anti-

bodies that produced negative results, in fourand two cases, respectively, when all of theothers were positive. However, D33 did havethe second highest score for 46% or morepositively labelled cells.The two main advantages of the EPOS

method with desmin antibodies are (1) in asmall biopsy sample there is a greater chanceof picking up positive cells and (2) the resultis obtained much faster. The routine methodtakes about four to fivehours from the timeappropriately dried sections are available,whereas the EPOS method takes about oneand a half hours. The latter method is alsomuch simpler, titrations are not required andso can be performed by someone with lessspecialised training. The time saving and thesaving of highly qualified staff offsets the extraexpense of the EPOS kits, which cost ap-proximately £2.00 per slide.We conclude that with the routine method

desmin clones labelled the same overall numberof rhabdomyosarcoma cases but the caseslabelled were different and therefore for thegreatest chance of accurately assessing themuscle of origin of a tumour, more than oneclone should be used. The EPOS method fordetecting desmin is fast and accurate with sig-nificantly more cells being labelled than withthe routine extravidin biotin technique. If only

537

on Novem

ber 28, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.48.6.535 on 1 June 1995. D

ownloaded from

Pollock, Rampling, Greenwald, Malone

one desmin clone is to be used with the routinemethod, we found D33 to be the best.

1 Small JV, Sobieszek A. Studies on the function and com-

position of the 10 nm (1 OOA) filaments of vertebratesmooth muscle. Cell Sci 1977;23:243-68.

2 Atlmannsberger M, Weber K, Droste R, Osborn M. Desminis a specific marker for rhabdomyosarcoma of human andrat origin. Anm _Pathol 1985;118:85-95.

3 Dias P, Kumar P, Marsden HB, Morris-Jones PH, Birch J,Swindell R, et al. Evaluation of desmin as a diagnosticand prognostic marker of childhood rhabdomyosarcomaand embryonal sarcomas. Br] Cancer 1987;56:361-5.

4 Ip W, Heuser JE. Subunit structure of desmin and vimentinprotofilaments and how they assemble into intermediatefilaments. Ann N YAcad Sci 1985;455:185-99.

5 Debus E, Weber K, Osborn M. Monoclonal antibodies todesmin, the muscle-specific intermediate filament protein.EMBO 1983;2:2305-12.

6 Wick MR. Antibodies to desmin in diagnostic pathology.In: Wick MR, Siegal GP, eds. Monoclonal antibodies indiagnostic imninnunohistochennsrst. New York: Marcel DekkerInc., 1988:93-114.

7 Carter RL, Jameson CF, Philp ER, Pinkerton CR. Com-parative phenotypes in rhabdomyosarcoma and de-veloping skeletal muscle. Histopathology 1990;17:301-9.

8 Malone M. Soft tissue tumours in childhood. Histopathologv1993;23:203-16.

9 Miettenen M, Lehto V-P, Badlev RA, Virtanen I. Alveolarrhabdomyosarcoma: demonstration of the muscle type ofintermediate filament, desmin, as a diagnostic aid. An 7

Pathol 1982;108:246-51.10 Tsokos M. The role of immunocvtochemistrv in the diag-

nosis of rhabdomyosarcoma. Arch Pathol Lab Med 1986;110:776-8.

1 1 Parham DM, Webber B, Holt H, Williams WK, Maurer H.Immunohistochemical study of childhood rhab-domyosarcomas and related neoplasms. Results of an

Intergroup Rhabdomyosarcoma study project. Cancer199 1;67:3072-80.

12 Skalli 0, Schurch W, Seemayer T, Lagace R, MontandonD, Pittet B, et al. Myofibroblasts from diverse pathologicalsettings are heterogeneous in their content of actin iso-forms and intermediate filament proteins. Lab Invest 1989;60:275-85.

13 Ordonex NG. Antidesmin antibodies. Ai 7 C/rn Par/wi

1990;93:430- 1.14 Truong LD, Rangdaeng S, Cagle P, Ro JY, Hawkins H,

Font RL. The diagnostic utility of desmin: a study of 584cases and review of the literature. Amii 7 Clin Pcathol 1990;93:305-14.

15 Parham DM, Dias P, Kelly DR, Rutledge JC, Houghton P.Desmin positivity in primitive neuroectodermal tumoursof childhood. An _7 Sinrg Pathol 1992;16:483-92.

538

on Novem

ber 28, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.48.6.535 on 1 June 1995. D

ownloaded from